#8320: Build fails due to Haddock error in ghc-prim:GHC.Types ----------------------------------------+--------------------------- Reporter: guest | Owner: Type: bug | Status: new Priority: highest | Milestone: 7.8.1 Component: Build System | Version: 7.7 Resolution: | Keywords: Operating System: Linux | Architecture: x86 Type of failure: Building GHC failed | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: ----------------------------------------+--------------------------- Comment (by darchon): It's still failing with the same error (latest commit: d5075dcfecbab1b6e3dbb7f8347124923796d46b) My complete build log: https://www.dropbox.com/s/c8ndh4x2jwtm4n1/build.log [[BR]] The interesting part: http://lpaste.net/93924 [[BR]] My build.mk: http://lpaste.net/93925 (`perf` build without profiling libs) The command that I used to built are: {{{ make maintainer-clean ./sync-all pull perl boot ./configure make }}} However, I did patch the intree haddock (`utils/haddock`) to get the extra interface-tracing information. You can find the patch I applied here: https://gist.github.com/christiaanb/6837931. The first part of the patch fixes a bug related to staticFlags (#8276), the second part adds the interface-tracing DynFlag. Anyhow, I looked at the part of the code that prints the error message, and it seems to print it when you want to read the interface file of a module in the HomePackage; an interface of a module it expects to be already read/loaded. The reason it got to that part of the code is because the interface info for the wired-in thing ([] in this case) was neither in the HomePackageTable (HPT), nor in the PIT. But since I don't know what either of those things are, nor what assertions should hold for them in relation to wired-in things, I couldn't analyse the bug any further. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8320#comment:25 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler